US3832683A - Character-identification device - Google Patents

Character-identification device Download PDF

Info

Publication number
US3832683A
US3832683A US00371902A US37190273A US3832683A US 3832683 A US3832683 A US 3832683A US 00371902 A US00371902 A US 00371902A US 37190273 A US37190273 A US 37190273A US 3832683 A US3832683 A US 3832683A
Authority
US
United States
Prior art keywords
character
circuit
identification
operator
decision
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime
Application number
US00371902A
Other languages
English (en)
Inventor
M Nadler
C Masson
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Bull SAS
Original Assignee
Honeywell Bull Sa
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Honeywell Bull Sa filed Critical Honeywell Bull Sa
Application granted granted Critical
Publication of US3832683A publication Critical patent/US3832683A/en
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/24—Character recognition characterised by the processing or recognition method
    • G06V30/248—Character recognition characterised by the processing or recognition method involving plural approaches, e.g. verification by template match; Resolving confusion among similar patterns, e.g. "O" versus "Q"
    • G06V30/2504—Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches

Definitions

  • This device includes a number of identification circuits equal to the number of character patterns which are validated by a validating and global a set of discrimination circuits included in each identification circuit,
  • the object of the present invention is a characteridentification device intended to be integrated into a recognition system for printed characters.
  • the known recognition systems consist of a device for point-by-point' scanning of the surface on which characters to be recognized are printed, for example by means of reading media, such as those described in a co-pending application Ser. No. 289,392, filed Sept. 15, 1972, entitled: "Read Head for an Optical Character-Recognition Systemwhich is assigned to assignee of this application.
  • This scanning device is followed by a device for analysis of the elements in which the scanned surface is divided, which makes it possible to send the resulting coded information to a shift register in matrix form, called: a character store.
  • This memory is conceived in such a way that each stored bit of information appears successively in all possible positions, shifted vertically and horizontally, before being rejected. This shift is performed in synchronization with the analysis and permits one to enter, for a given character, successively in the memory different elements of this character and the complete character.
  • the instant at which the complete character and certain of its elements are lodged in the memory may be determined.
  • These means enable a logical identification device, connected with different storing-position systems of the memory, to decide the identity of a character to be recognized, and of a character. pattern of a given set.
  • Identification devices were developed on the basis of statistical analyses which make it possible to keep in evidence the most frequent defects in such a manner that a character can be identified in spite of the existence of the defects.
  • Such a device comprises a certain number of logical circuits, calledmasks, each connected with a given group of positions of the character memory to detect the presence of certain information combinations in the memory and to emit a signal in the case of detection. It is characterized by a system of counters, each connected with a different mask circuit for counting the detection signals by these masks and by decision elements for the comparison of the contents of the counters and for identification of a character by the counter containing the highest count. The decision is only made under the condition that said count be at least equal to a predetermined minimum count and that in the other counters, no count exists whose difference from the highest count be below a predetermined minimum difference.
  • a character may be considered a group of printed elements covering determined adjoining zones, each element consisting of a set of printing points which are registered in the different positions of the character memory.
  • a character imperfectly printed includes then, at least one element, not printed or partially printed on the corresponding given zone, or printed on a wider extended zone than the given zone, when there are smudges.
  • the same character may comprise, for example, given zones, not printed by mistake, and too widely extended printed zones.
  • an identification device such as that mentioned above, working with integrationlogic, risks the elimination of a character exhibiting numerous missing print points, but whose present points arecharacteristic without ambiguity, with another character. It could also happen that two counts are higher than the others without presenting a difference between them so as not to create an ambiguity, although the character to be recognized has sufficient specific elements for its identification. A higher count than the others could also entail a mistake in interpretation, if it is caused by ridges.
  • One of the objects of the present invention is the very precise identification of a character exhibiting a printing defect, which are the cause of ambiguities not eliminated by the known identification devices.
  • the characteridentification device for N character-pattern comprising N identification circuits that operate with masks, a circuit for validation and for complete centering, and a decision circuit, is characterized in that:
  • the identification device receives from a character memory 10 an image x of the character to be recognized;
  • each identification circuit being associated with a character pattern is available for the identification of the image x with that character;
  • one character n is one of the N model characters and one character i is any of the (N-l) characters other than n;
  • the character n is statistically characterized by the elements El (n/i) which are specific for it in relation to the character i and which differentiate it completely from the character i;
  • an identification circuit n includes a mask circuit 21 and at the most (N-l) discrimination circuits of which one is a discrimination circuit i.
  • the mask circuit 21 includes a first mask circuit Ml, validated by the validation circuit, permitting a first level of decision, and the validation of the second mask M2 for a second level of decision;
  • the discrimination circuit 1' includes a first operator P1 (i),,, associated, on the one hand, with the elements Ell (n/i) and, on the other hand, with the elements El (i/n), a second operator P2 (i associated with the elements E2 (i/n), and a logic gate L whose two inputs are linked with a first output of the operator Pl (i and with the output of the operator P2 (i respectively, the second output of the operator P!
  • the operator PI serving to evaluate if the elements of the image x to be attributed, partly, to the elements El (n/i) and, partially, to the elements El (i/n) allow a choice by comparison between the characters n and i, or to indicate an ambiguity, and in the case in which the operator P! (i does not choose the character 1', the operator P2 (l), serves to indicate if the elements E2 (i/n) are not elements of the image x so that a de cision, at least on one level, may be made in certain fashion in favor of the character n, in relation to the character i;
  • FIG. I is a representation of several contents of the character memory with which the identification device is connected, according to the invention.
  • FIG. 2 is a generaldiagram of the identification device according to the invention.
  • FIG. 3 is a general diagram of an identification circuit being a part of the identification device according to the invention.
  • FIG. 4 is a partial diagram of a first example of embodiment of the identification circuit with two levels of the device of the invention
  • FIG. 5 is a partial diagram of a second example of embodiment of an identification circuit with two levels of the device according to the invention.
  • FIG. 6 is a partial diagram of a design of an identification circuit with one level of the device of the invention.
  • FIG. 7 is a decision table made on two levels by an identification circuit such as the one of FIG. 4.
  • FIG. 8 represents characters and their specific elements, taken into consideration by the operators of certain discrimination circuits of the device of the invention.
  • FIG. 9 shows an example of an operator having an effect on the specific elements of two model characters, and geometrical representation of the responses which it gives as a function of the elements present in the image of the character to be identified;
  • FIG. 10 depicts some examples of deformed characters recognized by the operator of FIG. 9;
  • FIG. 11 shows two examples of operators and the specific elements of two characters, one in relation to the other which these operators influence;
  • FIG. 12 presents some examples of deformed characters recognized by the operator of FIG. 11.
  • FIG. I3 presents some examples of deformed characters recognized by the different operators of the device according to the invention.
  • FIG. ll represents the images of different deformed characters, such as are recorded in the memory.
  • the characters of FIGS. Ia and lb each include three bars which are common elements of the 3 and E1.
  • the element specific for character 3 in relation to the character III is a central horizontal bar, while the element specific to the character [1 in relation to the character El is a vertical left bar.
  • the character exhibited in FIG. He displays no ambiguity for it includes no bar or section of a bar that could belong to another character than the character C].
  • certain important changes of the outline of the characters do not cause any risk of confusion.
  • other alterations produce an ambiguity which cannot be eliminated except by utilizing the highly specific characteristics of the outline, which are for this reason of utmost importance.
  • the identification device has been conceived by bearing in mind the statistically chosen elements so that no character is rejected, even when greatly changed, to the extent to which it remains in its outline characteristics which do not render it ambiguous.
  • the statistically representative elements which are determined according to a procedure and criteria depending on the special problem to be solved are not described here.
  • the examples offered hereinafter define the essential characteristics of the device according to the invention, and its possibilities.
  • the model characters number N and are designated by the figures from 1 to N including particularly the characters marked n, i and j.
  • the identification device illustrated in FIG. 2 includes N identification circuits, each associated with one of the N characters, respectively, whose circuits ll, 2, n, N, in parallel between a character memory and two register systems 11 and I2 enclosed in'a decision circuit 13, a validation circuit and of complete centering 14, whose input is connected with a group of positions of the character memory 10, with an output 15, linked in parallel with N identification circuits and an output 16 tied to the decision circuit 13.
  • the segmentation consists of the generation of an end of character signal, transmitted from the output 16 to the decision circuit 13, which initiates the working of the contents of the registers in the output of each identification circuit, then their reduction to zero.
  • the complete centering corresponds with the system of position for which the image present in the character memory 10 may be considered as correctly centered for each character.
  • the validation consists in defining the positions for which the content of the character memory 10 is taken into consideration by the group of N identification circuits which receive a signal from the output 15 of the validation circuit 14.
  • the two decisions made on two levels by the identification circuit n of FIG. 2 are stored in the register 17 and 18 of the systems I1 and 12, respectively.
  • each decision made on a level by one of the identification circuits is stored in a register.
  • the decision circuit 13 makes it possible to recognize a character identified by a single identification circuit, or to reject it when it is identified by several identification circuits, or when it is not identified by any.
  • the identification circuit n of FIG. 2 is also shown in FIG. 3 with its connections to the character memory 10, to the output 15 of the validation circuit 14 and to the registers 17 and 18 of the decision circuit 13.
  • This identification circuit includes at the most (N-l) discrimination circuits, such as the circuits 1, i and (N-l) of FIG. 3.
  • the responses which these discrimination circuits give to the information received from the character memory 10 are transmitted by two outputs of each of them, such as the outputs l9 and 20 of the discrimination circuit i, to a mask circuit 21.
  • the two outputs of this mask circuit are linked to the registers 17 and 18;, respectively.
  • the two decisions made by the identification circuit n are the result of responses given by all discrimination circuits, such as the circuits 1, i and (N-l).
  • FIG. 41 is a first example of embodiment of the identification circuit n with two levels of decision connected with two registers 17 and 18 of the decision circuit 13 of FIGS. 2 and 3.
  • Each of the discrimination circuits i and j is, connected with a given system of positions of the character memory 10 (the corresponding connections are not shown in FIGS. 4, 5 and 6 so as not to overcrowd them) and two outputs of each are linked to the mask circuit 21.
  • the discrimination circuit i includes two operators Pll (i),, and P2 (1') and a gate L.
  • the discrimination circuit j includes two operators Pl (1') and P (1')", and a gate L.
  • the mask circuit T3 includes two masks MI and M2 which are the AND gates whose outputs are linked to the registers 17 and 18, respectively.
  • the mask M1 is validated by the validation circuit 14, while the mask M2 is validated by the mask MT.
  • the operators P1 (1') P2 (i),,, Pl (1') and P2 (1') are designed initiated from the following elements: E1 (n/i), El (i/n), E2 (i/n), E1 (n/j), El (j/n), E2 (i/n).
  • the elements El (n/j) are the specific elements of the character n in relation to the character j.
  • the elements El (j/n) and E2 (j/n) are the proper and specific elements j in relation to the character n.
  • the connections of the discrimination circuits i and j with the character memory 10 are performed by the operators P1 (i),,, P2 (i Pl (j),,, P2 (1')
  • the presence of a signal at the output 22 of the operator Pl (i), indicates that the elements of the analyzed image x, to be attributed to the elements E1 (i/n), overbalance those, to be attributed to the elements El (n/i), i.e., decide for the character 1'.
  • the presence of a signal at the output 23 of the operator Pl (i) indicates a decision in favor of character n.
  • the operator Pl (i) indicates at its output 24 an option in favor of the character j and at its output 25 an option in favor of the character n.
  • the absence of a signal at the outputs 24 and 25 indicates an ambiguity, not removed by the operator Pl (j),,.
  • the presence of a signal at the output 26 of the operator P2 (i) indicates that elements E2 (i/n) are not present in the analyzed image x.
  • the presence of a signal at the output 27 of the operator P2 (1') indicates that the elements E2 U/n) are not present in the image x.
  • the gates L of the discrimination circuits i and j of FIG. 4 are OR gates. Thus, a signal appears at the output 30 of the gate L of the circuit i, when at least one of the two operators Pl (1'), and P2 (i), emits a signal.
  • the signals stemming from the gates L are only taken into account by the mask M2 when it is validated by the output 29 of the mask M1, i.e., when no operator, such as P1 (i decides in favor of a character other than n.
  • the mask M2 only emits a signal at the output when all of the operators Pl, such as Pl (i),, and Pl (j),,, opt in favor of a character n, or if certain of these operators are neutral, when the corresponding operators P2, such as P2 (i n and P2 (1 Opt in favor of the character n.
  • Pl (1') represented in FIG. 4 is an operator, termed divisive to the extent to which the characters n and i have common elements allowing one to consider by the same centering, the element El (n/i) and El (i/n).
  • the terminals 22 and 23 may be tied to the discrimination circuit n of the identification circuit i, similar to the discrimination circuit n.
  • the introduction of operators represents a reduction of logic circuits in the identification device according to the invention.
  • the operator Pl (1') represented in FIG. 4 is an operator, termed indivisible to the extent to which the characters n and j have no common elements sufficient for the same centering to consider the elements El (n/j) and E (i/n). This applies for example to the characters and A variation of the design of the identification circuit n is presented in FIG. 5.
  • the outputs 23 and 25 of the operatorsPl (i), and Pl (1') respectively, are linked to logic majority circuit 32 as well as the corresponding outputs of the other similar operators of the identification circuit n.
  • the output 33 of this majority circuit 32 is tied to one of the inputs of the mask Ml.
  • the threshhold of the circuit 32 being defined for a given number s of signals originating from operators, such as Pl (i),, and P1 (1')"
  • the mask Ml is not validated except when the threshhold is reached, i.e., for a minimum number s of discrimination circuits having opted in favor of the character n.
  • the introduction of this circuit 32 then entails a stricter decision by the masks Ml and M2 than the one made by the identification circuit n of FIG. 4.
  • FIG. 6 A simplifieddesign of the identification circuit n and of the decision circuit 13 of FIG. 4 is presented in FIG. 6.
  • the register 17 is omitted so that the mask Ml makes no decision on a first level, but functions only to validate the mask M2 so that a single decision be made by the latter and recorded in the register 18.
  • the decision circuit 13 (in FIG. 2) then, only includes the group of registers 12.
  • the decisions made by the identification device according to FIG. 6 are stricter than those made by the device of FIG. 4.
  • FIG. '7 is a summary table, presented as an example, of the decisions made on two levels by the identification circuits n and i (without a majority logical circuit) as a function of the indications given by their discrimination circuits i and n.
  • the parameters listed on this table correspond with the following data:
  • D (i) corresponds with the option of P2 (11),- (of the discrimination circuit n of the identification circuit i) in favor'of the character i.
  • V corresponds with one of three conditions:
  • FIG. 7 shows particularly that the conditions A, a being carried out for the identification circuit n, that the conditions B, c being carried out for the identification circuit i,
  • the operator Pl (i opts in favor of the character n, while the operator Pl (n), opts in favor of the character i, or the operator Pl (n),- is neutral while the operator P2 (n), opts in favor of the operator Pl (n), is neutral and the operator P2 (n),- opts in favor of the character i.
  • There is a single decision in favor of the character i if the operator Pl (n), opting in favor of the character 1; the operator Pl (i), is neutral and the operator P2 7 (1') does not opt in favor of the character n.
  • the decisions enclosed in dotted frames in FIG. 7 are those made on the second level by the mask M2, called hard level contrary to the soft level" which denotes the first level. This term is justified due to the fact that a decision can only be taken on the second level to the extent to which a'decision on the first level is made in favor of the character with which the identification cir-' cuit is associated. It is also justified due to the fact that a decision on the second level causes the intervention of the operators P2 having the local and specific characteristics of the characters with which they are asso- I) It consults all soft-registers for a decision on the soft level:
  • FIG. 8 presents for exemplary purposes the common elements represented by the white parts) and the specific elements (represented by the cross-hatched parts), defined statisticall respectively for the characters If] and [1 S and [I], %and 5 to produce the operators P1 the equivalents of the operators P1 '(i),, and P1 (j),, of FIGS. 4, 5 and 6.
  • the specific elements when considered vary in number from character to character, in extension and in position in the bars of which they are part.
  • the choice of these elements is basically made in such a way as to select the most characteristic printed zones of a character located between two common zones with another character.
  • the zones surrounded by plain strokes in FIG. 9a are elements common to the characters E and [1, while the zones bordered by dashes are the specific elements of these characters.
  • the zone E is a specific element of character 5 and the zones C and D are specific elements of character El.
  • the operator Pl E [I] equivalent to the operator Pl (1') of FIGS. 4, 5 and 6, is connected with a group of positions of the character memory It) so that it can see the elements C, D and E when the analyzed image x is framed in the memory.
  • a detector element not shown here, consisting of operational amplifiers, for example, evaluates the elements present in the image x, to be attributed, partially, to the specific element E and, partially, to the specific elements C and D, and indicates by a binary code the absence, or the partial or total presence, of each of the three elements C, D and E.
  • This code may, for example, be the following.
  • the comparison of the result S of said addition on two adjustable threshholds 61 2 and 62 4, in two differential amplifiers 35 and 36 provides the indication of an option in favor of the character 5 if S 2 and in favor of the character I] if S 3 4 after reversal of the signal in the output of the amplifier 36.
  • FIG. shows a geometrical representation in 3- dimensional space, of the options taken by the operator Pl E) El as a function of the variables (E1, E2), (Cll, C2), (D1, D2) and S, specific for the image x.
  • the value S 1, resulting from the values: E1 1, E2 1 with CT or D1 1 and CI 0 corresponds with the complete presence of the element E and the partial presence of one of the two elements C and D.
  • the value S 1, resulting from the values: E1 1, E2 0 with C1 C2 D1 D2 0 corresponds with the partial presence only of the element E.
  • An option in favor of the character E is well justified in these three cases. In the same manner, one may observe in FIG. 90 that an option in favor of the character D is taken in the following cases:
  • the elements C and D are fully present, while the element E is absent (S 6), or partially present (S 5) or totally present (S 4); 2. one of the elements C and D is fully present,
  • FIG. 10 offers examples for which the operator Pl E is capable of contributing to the identification of the deformed characters.
  • the characters of FIGS. Illa and 10b are recognized as a the characters of FIGS. 10c, 10d and We are recognized as I], in the same way as the character of FIG. 10f (by the predominance of the totally present elements C and D).
  • FIGS. 11d and 11h are operators P2 q 3 and P2 3-) 5 similar'to the operators P2 (i), and P2 (j),, of FIGS. 4, 5 and 6.
  • the cross-hatched zones F, G, H and Iof the FIGS. Ila, 11b, llc, 11d, Ile, 11f and 11g represent groups or marker positions of the character memory 10 by relation to which the horizontal bars of the image x (if they exist) may be framed when each of these zones occupies the center of a bar. This is true of the zones F and I placed respectively in the center of the two horizontal bars of the image of the character of FIG. 11a. In FIG. Ilb the zone F is centered, while the zone I is centered in the FIG.
  • FIG. Me the zones G and H are centered, while in the FIGS. Ilf and Hg it is the zones G and H which are centered, respectively.
  • the specific elements defining the operators P2 q 3 and P2 3 q of the Ill FIGS. lid and 11h are the intervals of the horizontal bars determined by the relative positions determined by the zones F, G, H and I, among each other.
  • the image of a character is recognized as a 3 if the zones F, G, H, I do not occupy relative positions corresponding with the horizontal bars of a 3 the zone F centered in the upper bar with the zone H in the interior of the lower bar (as in FIG. Ilb), or the zone I centered in the lower bar with the zone G inside the upper bar.
  • the operator P2 3 of the FIG. I Id is then defined by the following conditions:
  • the detectors permitting the analysis of the location of the different zones in relation to said bars are effected by known means.
  • the AND gate 37 of FIG. 11d tied to two of these detectors, emits a signal at its output when the conditions (F hidden) and (H white) are accomplished.
  • the AND gate 38 tied to two other detectors, emits a signal at its output when the conditions (I framed) and (G white) are achieved.
  • the OR gate 39 receiving at least one signal from these gates 37 and 38, indicates at its output an option in favor of the character 3 of the discrimination circuit 3 contained in the identification circuit 3.
  • the operator P2 3 q recognizes the image of a character as 3 if the zones F, G, H and I do not occupy relative positions corresponding with the horizontal bars of an 3 zone G centered in the upper bar with zone'l inside the lower bar (as in FIG. 11)), or zone H centered in the lower bar with zone F inside the upper bar (as in FIG. 11g).
  • the operator P2. 3 q comprises two AND gates 40 and 41 emitting a signal at their output when the following conditions are met: zone H framed with zone F outside the upper bar (F white) and zone G framed with the zone I outside the lower bar (1 white).
  • FIGS. 12a, ll2b and 120 present three examples of character images which a discrimination circuit :1 (contained in the identification circuit 3 of a device according to the invention) is capable of recognizing by its operators P1 3 a and P2 3 q similar to those of FIGS. 9b and lid, respectively.
  • a discrimination circuit :1 contained in the identification circuit 3 of a device according to the invention
  • FIGS. 12d, l2e and 12f give three other examples of character images which a discrimination circuit 3 (contained in the identification circuit 3 of a device according to the invention) is capable of recognizing by its operators Pl 3 and P2 (3 similar to those of FIGS. 9b and 11h, respectively.
  • the operator Pl 3 makes a decision in favor of the character 3 in the case of FIG. 12d by considering the same elements I and K as the operator Pl q 3
  • the images of the characters of the FIGS. 12a and 12f cause an ambiguity for the operator Pl 3 In this case it is the operator P2 3 q as that of FIG. 1 lb which makes a decision in favor of the character 3 by evaluation of the interval between the horizontal bars of these two images.
  • FIGS. 11d and Elk affect the specific elements of the characters each including two horizontal bars whose deviation is different as the pair of characters 3 and
  • the specific elements of the characters 3 and El making it possible to produce the operators P2 3 )[j and P2 ([1); are, for example, the beginnings of bars, such as the localized zones 43 and representing them in FIGS. 13b, 130, 13e, 13f, 13g, 13h and 131'.
  • FIG. 13a and 13d illustrate, as in FIG. 12, in which cases, and when, the operators P2 intervene with the operators P1 (of the same discrimination circuits) for decisions on the soft and hard level.
  • a single decision is made on the soft" level for the recognition of the character U by option of the operator Pll 3 of the identification circuit II], in favor of I].
  • a multiple decision is first made on the fsoft level by the operators Pl ([1); and P1 3 of the identification circuits 3 and [I which remain neutral.
  • the operator P2 (B) 3 belonging to the discrimination circuit C1 of identification circuit 3 then intervenes with the specific elements 43 for a rejection of the character 3 on ahard level by the identification circuit 3.
  • the operator P2 3 belonging to the discrimination circuit 3 of the identification [I] intervenes with the specific elements 44 for a single decision on the hard level in favor of the character E] by the identification circuit C].
  • the character of FIG. is recognized as E] on a soft" level of decision by intervention of the operator Pl 3 j of the identification circuit (:1, without the operator P2 3 being consulted.
  • the character of FIG. 13d is recognized as 3 on a soft level of decision of the identification circuit 3, without the operator P2 ([1) 3 being consulted.
  • the character of the FIG. l3e causes an ambiguity for the operators PI (El); and P1 3 )U of the identification circuits 3 and CI.
  • the consultation of the op erators P2 ([1); and P2 3 ) ⁇ 1 produces a hard decision in favor of 3 by the identification circuit 3 and a rejection on a hard level by identification circuit E].
  • the characters of FIGS. 13g and 13h cause ambiguities for the operators Pl ([1) 3 and P1 3 )[j In the case of FIG.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Character Discrimination (AREA)
  • Character Input (AREA)
US00371902A 1972-06-30 1973-06-20 Character-identification device Expired - Lifetime US3832683A (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
FR7223669A FR2191788A5 (de) 1972-06-30 1972-06-30

Publications (1)

Publication Number Publication Date
US3832683A true US3832683A (en) 1974-08-27

Family

ID=9101117

Family Applications (1)

Application Number Title Priority Date Filing Date
US00371902A Expired - Lifetime US3832683A (en) 1972-06-30 1973-06-20 Character-identification device

Country Status (4)

Country Link
US (1) US3832683A (de)
DE (1) DE2333202A1 (de)
FR (1) FR2191788A5 (de)
GB (1) GB1440517A (de)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4030068A (en) * 1976-01-12 1977-06-14 Decision Data Computer Corporation Optical character recognition system
US4075605A (en) * 1974-09-13 1978-02-21 Recognition Equipment Incorporated Character recognition unit
EP0036150A2 (de) * 1980-03-13 1981-09-23 Kabushiki Kaisha Toshiba System zur Mustererkennung nach dem Mehrfach-Ähnlichkeitsverfahren
US5142590A (en) * 1985-11-27 1992-08-25 Trustees Of Boston University Pattern recognition system
US5257323A (en) * 1991-05-29 1993-10-26 Canon Kabushiki Kaisha Selection agent for a symbol determination system with multiple character recognition processors
US8285791B2 (en) 2001-03-27 2012-10-09 Wireless Recognition Technologies Llc Method and apparatus for sharing information using a handheld device

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB1476880A (en) * 1974-06-24 1977-06-16 Ibm Data processing apparatus

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3588823A (en) * 1968-03-28 1971-06-28 Ibm Mutual information derived tree structure in an adaptive pattern recognition system
US3688267A (en) * 1969-11-05 1972-08-29 Taizo Iijima Pattern identification systems operating by the multiple similarity method
US3790955A (en) * 1970-05-27 1974-02-05 Klemt Kg Arthur Raster process for classifying characters

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3588823A (en) * 1968-03-28 1971-06-28 Ibm Mutual information derived tree structure in an adaptive pattern recognition system
US3688267A (en) * 1969-11-05 1972-08-29 Taizo Iijima Pattern identification systems operating by the multiple similarity method
US3790955A (en) * 1970-05-27 1974-02-05 Klemt Kg Arthur Raster process for classifying characters

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4075605A (en) * 1974-09-13 1978-02-21 Recognition Equipment Incorporated Character recognition unit
US4030068A (en) * 1976-01-12 1977-06-14 Decision Data Computer Corporation Optical character recognition system
EP0036150A2 (de) * 1980-03-13 1981-09-23 Kabushiki Kaisha Toshiba System zur Mustererkennung nach dem Mehrfach-Ähnlichkeitsverfahren
EP0036150A3 (en) * 1980-03-13 1983-11-09 Tokyo Shibaura Denki Kabushiki Kaisha Pattern recognition system operating by the multiple similarity method
US5142590A (en) * 1985-11-27 1992-08-25 Trustees Of Boston University Pattern recognition system
US5257323A (en) * 1991-05-29 1993-10-26 Canon Kabushiki Kaisha Selection agent for a symbol determination system with multiple character recognition processors
US8285791B2 (en) 2001-03-27 2012-10-09 Wireless Recognition Technologies Llc Method and apparatus for sharing information using a handheld device

Also Published As

Publication number Publication date
FR2191788A5 (de) 1974-02-01
DE2333202A1 (de) 1974-01-24
GB1440517A (en) 1976-06-23

Similar Documents

Publication Publication Date Title
US4408342A (en) Method for recognizing a machine encoded character
US3136976A (en) Method for the automatic recognition of characters, in particular writing characters
US4748317A (en) Optical reader
US4015240A (en) Pattern recognition apparatus
US3918028A (en) Hand held optical reader
US5185812A (en) Optical pattern inspection system
US3541511A (en) Apparatus for recognising a pattern
JP2011227636A (ja) 文字列認識装置および文字列認識方法
US3832683A (en) Character-identification device
JPH01502371A (ja) ナンバープレート読取装置
JPH0587871B2 (de)
JPH0467234B2 (de)
GB1390404A (en) Article identification apparatus
US4087790A (en) Character presence processor
CN107067533A (zh) 一种纸币鉴别的方法及装置
US3652992A (en) Method and apparatus for quantizing a character or test pattern preferably for the purpose of gaining control data for electronic photo composition
JPH0436433B2 (de)
US3833882A (en) Character reading system
KR102094234B1 (ko) 지폐 일련번호 인식 장치 및 방법
US3177469A (en) Character recognition
US3201751A (en) Optical character reading machine with a photocell mosaic examining device
US3571793A (en) Character recognition systems
GB1020647A (en) Character reader
US3104370A (en) Recognition systems using assertions and negations
KR890003689B1 (ko) 광학적 판독장치